本文探討金融資產報酬為非常態分配時，在最適避險策略的效 果。我們首度採用極小化student-t 分配之風險值及條件風險值來估 算避險比率。以英國金融時報100 指數、台灣加權股價指數、黃金 與西德州原油為現貨及其對應之期貨為避險工具， 探討極小化 Student-t 風險值與極小化Student-t 條件風險值之避險效果，是否較 極小化變異數、極小化Cornish-Fisher (Cornish and Fisher, 1937) 風 險值、極小化Cornish-Fisher 條件風險值模型等研究方法，提供較 佳之避險績效。實證結果顯示，極小化Student-t 風險值與極小化 Student-t 條件風險值避險法亦使避險組合分配產生高狹峰，但在風 險值與條件風險值減少程度上優於其他方法。最後，利用回溯測試 檢視各模型評估市場風險的能力。結果顯示只有極小化Student-t 風 險值法與極小化Student-t 條件風險值通過檢驗。代表極小化 Student-t 風險值與極小化Student-t 條件風險值法不但在避險績效上 優於其他方法並且能精準地預測風險值。

This paper evaluates the non-normality effect of financial asset returns on the optimal hedge ratio. We adopt a new method of estimating the minimum value at risk and the minimum conditional value at risk hedge ratios based on the student-t distribution. Using spot and futures returns for the FTSE 100, Taiwan Weighted Stock Indices, Gold, and WTI Crude Oil, we examine whether the new approach works better than the minimum-variance hedging, minimum value at risk, and minimum conditional value at risk approaches based on the Cornish- Fisher expansion. Among the various models, the empirical results show that the minimum student-t value at risk and the minimum student-t conditional value at risk techniques present more efficient results. Moreover, through backtesting, the empirical results offer evidence that the new approaches accurately evaluate the market risk of hedging portfolios.